Role of Protein Interactions in Retina Development and Function

National Eye Institute

Investigators

Linked publications, trials & patents

Abstract

A study on the characterization of photoreceptors of mice in which PNPLA2 gene was deleted was completed. Because phospholipids are the most abundant lipid class in the retina, we investigated the role of PEDF-R in photoreceptors by generating CRISPR Pnpla2 knock-out mouse lines in a retinal degeneration-free background. Pnpla2-/- mice had undetectable retinal Pnpla2 gene expression and PEDF-R protein levels as assayed by RT-PCR and immunofluorescence, respectively. The photoreceptors of mice deficient in PEDF-R had deformities as examined by histology and transmission electron microscopy. Pnpla2 knockdown diminished the phospholipase A2 enzymatic activity of PEDF-R in the retina. Lipidomic analyses revealed the accumulation of lysophosphatidyl choline-DHA and lysophosphatidyl ethanolamine-DHA in PEDF-R-deficient retinas, suggesting a possible causal link to photoreceptor dysfunction. Loss of PEDF-R decreased levels of rhodopsin, opsin, PKC, and synaptophysin relative to controls. Pnpla2-/- photoreceptors had surface-exposed phosphatidylserine, and their nuclei were TUNEL positive and condensed, revealing an apoptotic onset. Paralleling its structural defects, PEDF-R deficiency compromised photoreceptor function in vivo as indicated by the attenuation of photoreceptor a- and b-waves in Pnpla2-/- and Pnpla2+/- mice relative to controls as determined by electroretinography. Thus, ablation of PEDF-R in mice caused alteration in phospholipid composition associated with malformation and malperformance of photoreceptors. These findings identify PEDF-R as an important component for photoreceptor structure and function, highlighting its role in phospholipid metabolism for retinal survival and its consequences. To explore the regulatory role played by the PEDF/PEDF-R axis in photoreceptor structure and function, we continued characterizing photoreceptors deficient in PEDF and PEDF-R. CRISPR Pnpla2 knock-out mouse was crossbred with Serpinf1 null, and three genotypes were generated: wild type, Serpinf1+/- x Pnpla2-/- and Serpinf1-/- x Pnpla2-/-. Gene expression determined in dissected retinas by RT-PCR verified that Serpinf1+/- x Pnpla2-/- and Serpinf1-/- x Pnpla2-/- mice had decreased and undetectable retinal Serpinf1, respectively, relative to their Serpinf1+/+ x Pnpla2+/+ littermate controls and both had undetectable Pnpla2 expression levels. Mice were evaluated for electroretinography, fundoscopy and autofluorescence angiography. The electroretinography recordings showed that a- and b-waves of Serpinf1+/- x Pnpla2-/- and Serpinf1-/- x Pnpla2-/- mice were attenuated relative to controls. Whole eyes were enucleated and processed for histology, immunofluorescence, TUNEL, confocal microscopy, and transmission electron microscopy. The Serpinf1+/- x Pnpla2-/- mice had higher lipid droplet density in their RPE and Bruchs membrane, and the vasculature was more ectatic relative to controls. The Serpinf1-/- x Pnpla2-/- retinas had morphological abnormalities, thinner outer nuclear layers with TUNEL positive photoreceptor cells, as well as decreased levels of rhodopsin, opsin, protein kinase C alpha, and synaptophysin compared to controls. The biochemistry of the PEDF-PEDF-R protein complex continued to be investigated. Expression and purification of recombinant human PEDF and PEDF-R variants by mammalian and bacterial cells was optimized. To obtain large amounts of active recombinant PEDF-R, C-terminal truncated PEDF-R versions were subcloned into expression vectors and recombinant protein production and solubility were determined. One of them, PEDF-R(1-288) with N-terminal 6XHis and C-terminal 2XStrep-tags, was subcloned in the pESUMOstar vector. pGro7(groES-groEL), and pESUMOstar-PEDF-R(1-288) plasmids were co-transformed into Escherichia coli Rosetta(DE3), and the cells were cultured in LB media containing L (+) arabinose, chloramphenicol, and ampicillin at 37 C until the O.D. 600 nm reached 0.5-0.6. Protein expression was induced by addition of isopropyl-d-1-thiogalactopyranoside and continued culturing at 25 C overnight. The expression and solubility of PEDF-R(1-288) was determined after separation of soluble from particulate material by centrifugation, and by SDS-PAGE followed by Coomassie Blue staining and western blotting. Purification of the recombinant protein was optimized by HisTrap affinity column chromatography followed by StrepTrap affinity column chromatography using an automated AKTA system to obtain milligram amounts of highly purified protein. The size of the protein was characterized using size-exclusion column chromatography. Binding assays of PEDF-R and PEDF peptides were performed using MicroScale Thermophoresis (NCI, Ettore Appella). Phospholipase A2 activity of the recombinant protein was determined. The full-length human PEDF version with a single point alteration at amino acid position 105 (H105A) was produced at large scale by HEK.Ebna cells harboring an expression vector with the corresponding PEDF cDNA. The secreted proteins from the cells into the serum-free media were collected and used to purify the recombinant PEDF(H105A) protein by two steps of tandem cation- and anion-exchange column chromatography. Recombinant PEDF production and secretion were detected by SDS-PAGE followed by Coomassie blue staining and western blotting of the culturing media and fractions from the purification protocol. Stimulation of the phospholipase A2 activity of PEDF-R protein combined with 17-mer(H105A) peptide and PEDF(H105A) was determined. Derivatives of the 17-mer peptide were designed and synthesized (NHLBI, Rolf Swenson) to test stimulation of enzymatic PEDF-R activity. We continued investigating the regulation of the photoreceptor-specific transcription factor cone-rod homeobox (CRX) by PEDF in retinas of Serpinf1-/-, Serpinf1+/-, Serpinf1+/+ mice, rd10 and rd10 x Serpinf1-/- mice.